The present invention relates to a continuous process for the industrial preparation of propylene oxide from hydrogen peroxide and propylene.
Hitherto propylene oxide has been prepared on a large industrial scale by two processes exclusively, that is either according to the older process via propylene chlorohydrin or more recently with the aid of hydrocarbon peroxides.
The older chlorohydrin process has the disadvantage that undesirable chlorinated by-products and waste salts which pollute the environment are formed (DAS (German Published Specification) 1,543,174, column 2, lines 15 et seq.).
The more recent process, used industrially, for the preparation of propylene oxide via hydrocarbon peroxides, such as is described, for example, in USA Patent Specification 3,350,422, eliminates these considerable disadvantages of the chlorohydrin process. The reaction of propylene with a hydrocarbon peroxide ROOH can be illustrated by the equation (1) ##STR1##
It can be seen from equation (1) that in this reaction 1 mol of the alcohol ROH corresponding to the peroxide is always formed per 1 mol of propylene oxide formed. The hydrocarbon peroxide thus effects a transfer of oxygen so that, after the release of the peroxide oxygen, the corresponding alcohol is obtained as a co-product and frequently has to be removed as an undesired by-product. Accordingly, the possibilities for industrial use of such a process are limited, since the alcohol by-product cannot be utilised in every case.
In contrast, with the principle on which the process according to the invention for the preparation of propylene oxide from propylene and hydrogen peroxide is based, the desired end product is obtained, as is shown in equation (2), free from such by-products, which either have to be eliminated at considerable expense because of their environmental pollution properties or for which a suitable further use has to be found when they are obtained as co-products. ##STR2##
However, the desired objective is not achieved by direct reaction of propylene with aqueous hydrogen peroxide (USA Patent Specification 3,350,422, column 2, lines 42 - 44).
On the other hand, it is known to epoxidise propylene with the aid of a percarboxylic acid to give propylene oxide (Prileschayev, Ber. dtsch. chem. Ges. 42, 4811 (1909) and D. Swern "Organic Peroxides," Wiley Interscience 1971, volume 2, page 355-533, especially page 375-378 and page 397). In addition, it is known to obtain percarboxylic acids from carboxylic acids with the aid of hydrogen peroxide (German Patent 251,802 and, for example, D. Swern, loc. cit., 1970, volume 1, page 313-369 and page 428-439). These two partial steps are illustrated in the equations (3) and (4), in which R-COOH and R-COOOH represent a carboxylic acid and a percarboxylic acid respectively. ##STR3##
If the carboxylic acid obtained according to equation (4) is recycled into the reaction according to equation (3) to obtain percarboxylic acid, the overall equation (2) results for the reaction of hydrogen peroxide with propylene to give propylene oxide. A process of this type for the preparation of propylene oxide starting from hydrogen peroxide and propylene and using percarboxylic acids as the epoxidising agent has not hitherto been mastered in an industrially satisfactory manner and consequently has not yet been used on an industrial scale. In this connection it is stated, for example, in USA Patent Spec. No. 3,350,422 (column 1, line 65 to column 2, line 11):
"In light of the complexity and cost of the chlorohydrin route, workers have turned to other possible routes for the epoxidation of propylene and other olefins. One route which has proved successful insofar as being capable of actually producing at least limited yields of propylene oxide and other oxides is the peracid route. This route involves the formation of a peracid, such as peracetic acid, through the reaction of hydrogen peroxide with the organic acid and the epoxidation of an olefin with the peracid. The disadvantages of the peracid route also are such as to preclude significant commercialization. The peracids themselves are extremely hazardous to handle and give rise to severe operation problems. The reagents are expensive, corrosive, and nonregenerable, inasmuch as the hydrogen peroxide is lost as water. The composition of the peracid epoxidation mixture contains chemicals (H.sub.2 O, AcOH, and H.sub.2 SO.sub.4) which are highly reactive with the product epoxides, thus leading to many by-products (glycol, glycol monoester, glycol diester) which lower the overall efficiency. This problem becomes more severe with the less reactive olefins, in particular propylene."
In fact, all the processes hitherto known for the preparation of propylene oxide from hydrogen peroxide and propylene, which proceed via the intermediate stage of a percarboxylic acid as an oxygen transfer agent, lead only to unsatisfactory yields of propylene oxide and to considerable amounts of by-products, such as propylene glycol, propylene glycol monoester and propylene glycol diester. It has also not been possible satisfactorily to overcome the extremely difficult process problems, especially with regard to the isolation of the percarboxylic acid, which are caused by the explosion hazard of the percarboxylic acids.
In the case of the process according to DOS (German Published Specification) 1,618,625, which has been disclosed more recently, for the preparation of oxiranes from olefines and hydrogen peroxide with the aid of formic acid, the measures described there are also not adequate for an industrially satisfactory production of propylene oxide from hydrogen peroxide and propylene. For this process it is necessary for the reaction mixture to be substantially free from mineral acid and substantially anhydrous or to contain only a small amount of water (DOS (German Published Specification) 1,618,625, Claim 1). Thus, it is stated, for example, on page 3, final paragraph and page 4, first line, of DOS (German Published Specification) 1,618,625: "The use of an anhydrous reaction mixture is desired, but the preparation of solutions of performic acid having less than about 0.3% of water is neither simple nor economically tenable. The use of a reaction mixture which contains only a small amount of water is preferred." An amount of less than 20 g/l is mentioned as an appropriate water content and an amount of less than 10 g/l is mentioned as being a required water content in some cases. The freedom from mineral acid, which it is attempted to achieve in the process, is important since the catalysts required for the reaction of formic acid with hydrogen peroxide also catalyse the cleavage reaction of oxirane rings, in the present case the cleavage of propylene oxide (DOS (German Published Specification) 1,618,625, page, 5 lines 10-14). Accordingly, it would be most advantageous to use in the process a solution, which as far as possible is absolutely anhydrous and as far as possible is free from mineral acid, of performic acid in a hydrophobic solvent. These requirements, particularly with regard to the freedom from water, cannot be met in the processes known hitherto, since the preparation of a non-aqueous performic acid containing only 0.3% of water or less already comes up against the difficulties mentioned in DOS (German Published Specification) 1,618,625. Accordingly, the yield of propylene oxide which can be achieved, for example, according to the process of DOS (German Published Specification) 1,618,625, is only 85%, relative to the performic acid consumed (DOS (German Published Specification) 1,618,625, Example 3). However, since the performic acid solutions still have a relatively high content of free hydrogen peroxide, this being between 3 and 10 mol % of the performic acid according to Examples 1 and 2 of DOS (German Published Specification) 1,618,625, the yield of propylene oxide, relative to hydrogen peroxide employed, is even lower, since the hydrogen peroxide contained in the performic acid solution used as the epoxidising agent can not be recovered from the mixtures, containing propylene oxide, which are obtainable from the reaction with propylene. It is not possible to determine the accurate percentage figures for the final yield of propylene oxide, relative to hydrogen peroxide employed, from the data given in the examples; however, it is less than 50%.
A further disadvantage of the process of DOS (German Published Specification) 1,618,625 is that the formic acid used as the oxygen transfer agent is a special case amongst the carboxylic acids with regard to the question of corrosion also, which is always of considerable importance in reactions with lower carboxylic acids, because formic acid is even particularly corrosive towards stainless steels. It is precisely in a process in which sensitive peroxy compounds, such as hydrogen peroxide and percarboxylic acids, are used that corrosion of any type is extremely undesirable since, due to corrosion, heavy metal compounds which cause the decomposition of hydrogen peroxide and of the percarboxylic acid are carried into the reaction.
In another more recent process for the preparation of olefine oxides from olefine and hydrogen peroxide, an aromatic carboxylic acid, preferably benzoic acid, is used as the oxygen transfer agent (DOS (German Published Specification) 2,312,281). However, in this process the problem of obtaining the percarboxylic acid by reaction of hydrogen peroxide with an aromatic carboxylic acid has not been solved satisfactorily. That is to say, the reaction mixture, containing percarboxylic acid, which is obtainable must be diluted, for further working up, with ice water and cooled ammonium sulphate solution whilst maintaining a temperature of less than 25.degree. C and the unreacted hydrogen peroxide is then destroyed. (DOS (German Published Specification) 2,312,281, page 5, 2nd and 3rd paragraph). A further disadvantage of this process is that the rate of reaction of the aromatic percarboxylic acid with propylene is very low, since after a reaction time of 4 hours at a temperature of 28 to 30.degree. C only 66% of the percarboxylic acid are converted. The total yield of propylene oxide, relative to hydrogen peroxide employed, is apparently very small with this process. According to Example 1 of DOS (German Published Specification) 2,312,281, the final yield for propylene oxide, relative to hydrogen peroxide employed, is about 40%.
A further process which can be used to prepare propylene oxide is the process for the oxidation propylene described in DOS (German Published Specification) 1,917,031, in which propylene is reacted with an equilibrium mixture consisting of at least one carboxylic acid, hydrogen peroxide and water, in the absence of mineral acid and heavy metal ions, the amount of water present during the reaction being so regulated that at least one compound from the group comprising propylene oxide, propylene glycol and propylene glycol esters is obtained. When carrying out the process in practice, a hydrogen peroxide solution prepared by air oxidation of a secondary alcohol, for example isopropanol, is used as the starting material for the preparation of the equilibrium mixture to be employed in the process and is treated with a urea solution in order to form a urea/hydrogen peroxide adduct, which is mixed with an extracting solvent (an alkyl ketone, alkyl ester or alkyl orthphosphate, by which means the hydrogen peroxide is dissolved in the extracting solvent, urea being deposited, and subsequently at least part of the extracting solvent in the resulting hydrogen peroxide solution is mixed with the carboxylic acid, for example acetic acid, or replaced by this (DOS (German Published Specification) 1,917,031, page 3 and also Example 1). The oxidation of propylene then carried out using the equilibrium mixture leads to the formation of propylene oxide, propylene glycol and propylene glycol esters in varying amounts (loc. cit., page 4, lines 2 and 3). The ratio of propylene oxide to propylene glycol and propylene glycol esters is regulated by the amount of water and excess carboxylic acid which remains in the equilibrium mixture containing the percarboxylic acid (loc. cit., page 5, lines 6-8). When the process is intended to give propylene oxide as the main product, it is appropriately carried out, as can be seen from DOS (German Published Specification) 1,917,031, using only a slight excess of carboxylic acid, since, as is known, the presence of larger amounts of carboxylic acid easily leads to the formation of propylene glycol and the esters thereof and not to the formation of propylene oxide (loc. cit., page 6, lines 18 to 23). This in turn means that the rate of formation of the percarboxylic acid is reduced and this has an adverse effect on the economics of the process (loc. cit., page 7, line 1 to 4). Moreover, because of the absence of mineral acid, the rate of formation of the percarboxylic acid in this process is considerably lower at all molar ratios of hydrogen peroxide to carboxylic acid than when mineral acid is present. The effect of this is, of course, very particularly disadvantageous if the excess of carboxylic acid is small. The yields of propylene oxide, relative to hydrogen peroxide employed, achieved according to this process are small, especially because the unreacted hydrogen peroxide is not recovered and the unreacted percarboxylic acid is destroyed. Because of the lack of data, the yields of propylene oxide, relative to hydrogen peroxide employed, cannot be calculated accurately from the two illustrative examples of DOS (German Published Specification) 1,917,031. However, it can clearly be seen from the data of DOS (German published Specification) 1,917,031 that the peracetic acid solution prepared according to Example 1(a) must still have contained substantial amounts of free hydrogen peroxide, so that the yield of peracetic acid, relative to the amount of hydrogen peroxide employed, can have been about 69% in the most advantageous case. Accordingly, the yield of propylene oxide, relative to hydrogen peroxide employed, of course also falls considerably, to about 64% in Example 2(b,i).
Accordingly, it can be seen from the state of the art that it has not been possible to find a technically satisfactory solution, not only in respect of the process step for the preparation of the percarboxylic acid, but in particular also in respect of the subsequent reaction of the percarboxylic acid, for example as a non-aqueous solution, with propylene to give propylene oxide. Improvements in this reaction with regard to process engineering, such as have been described in British Patent Specification 1,105,261, German Patent Specification 1,216,306 and DOS (German Published Specification) 1,923,392, also have such great disadvantages that they cannot be used for carrying out the process on an industrial scale.
The basic assumption in British Patent Specification 1,105,261 is that only yields of 75%, relative to the percarboxylic acid, are possible when this reaction is carried out by mixing the reactants, for example by mixing propylene and peracetic acid (British Patent Specification 1,105,261, page 1, lines 20 - 24).
Now it is proposed in British Patent Specification 1,105,261 to use a series of closed reaction loops, in which mixing of reaction products with the starting substances is largely prevented, for carrying out the reaction of a non-aqueous peracetic acid solution with propylene. However, the proposed process is not adequate for an economical preparation of propylene oxide from propylene and a percarboxylic acid, since the yield of propylene oxide, relative to peracetic acid employed, is only 90% and 2.5 mol% of propylene glycol monoacetate and a further 2.5 mol% of other higher boiling by-products are formed (British Patent Specification 1,105,261, page 3, lines 60-68).
Even according to the process of German Patent Specification 1,216,306, by using coiled tubes of very precise dimensions for the reaction of propylene with peracetic acid, a yield of only 86% of theory is achieved. (German Patent Specification 1,216,306, column 8, line 33).
The process according to DOS (German Published Specification) 1,923,392 is intended to improve the rate of reaction and, at the same time, to prevent side reactions and secondary reactions, because, although the rate of reaction can be increased by simply carrying out the reaction under pressure, it has not been possible to prevent the occurrence of side reactions in this way (DOS (German Published Specification) 1,923,392, page 2, lines 14 - 18). According to the process of DOS (German Published Specification) 1,923,392, an attempt is then made to eliminate these disadvantages by using a reaction system consisting of a multiplicity of reaction zones (in practice a multi-stage bubble column). However, carrying out the reaction in this way means that, due to the requisite technically highly expensive procedure, a new and considerable disadvantage has to be accepted, because the process technology for the reaction of propylene with peracetic acid in heterogeneous phase (gaseous/liquid) is far more complicated than that for a reaction in homogeneous phase.